Large donations of free nets have allowed net ownership by households in Africa to increase markedly since 2000 , and ownership of long-lasting insecticidal nets (LLIN) in Ethiopia has increased dramatically in the last few years [2, 3]. After an initial peak in ownership and use shortly after distribution, net ownership and use have been observed to drop off in several countries [4, 5].
Clearly, net ownership is a necessary prerequisite for net use. However, whether or not a net owner will use a net every night, some nights, or not at all depends on complex multi-level interactions between individual characteristics, household characteristics, social and cultural factors, community-level factors, aspects of the physical environment and characteristics of the net itself.
In Ethiopia, a decline was observed in net use in households owning nets between two representative household surveys conducted approximately one year apart, and this did not appear to be associated with differences in sampling or any climatic or seasonal differences between the survey years . To investigate further the reasons for the decline, characteristics of nets that may be impacting their use in Ethiopia were studied. Between 2006 and 2007, the proportion of households owning at least one net increased dramatically from 19.6% to 65.6%, but the proportion of nets used the previous night in households owning nets decreased from 85.1% to 56.0% [3, 6]. In the net level analysis, factors independently associated in both surveys with reduced likelihood that a net would be used were: increasing net age, increasing damage of nets, increasing household net density (nets/person), and increasing altitude (> 2,000 m). Factors associated with increased likelihood of a net being used were: increasing wealth index (at both surveys), LLIN net type (in 2006), and household status of indoor residual spraying with insecticide (in 2007) .
At the individual level, factors influencing net use have been reported to include age and gender [5, 7–10], education, occupation/livelihood [11–13], degree of control over household decision-making , malaria knowledge, beliefs and risk perceptions [5, 15–21], perceived benefits and disadvantages of nets [5, 17, 18, 21], trust in health workers providing health education and LLINs , knowledge of appropriate net use/care practices, and net-hanging skills [23–25]. Household level determinants of net use include household size and composition , the number of children under five years of age (U5) in the household, intra-household sleeping arrangements [5, 13, 26], household structure and space [5, 7, 9, 18, 26, 27], household decision-making processes and power structures , and use of other vector control measures [14, 19, 28]. At the community level, social norms and values , cultural beliefs and practices [13, 20], mechanisms of LLIN distribution and distance to LLIN suppliers [14, 20], rumours about LLINs  and social support and pressure  all have the potential to influence net use by individuals and within households. For example, white nets may be associated with burial shrouds and death, and free nets may be regarded as toxic or even deliberately harmful to recipient groups . Important environmental factors include climate and temperature [7, 18], perceived mosquito density , availability and proximity of land for farming and grazing livestock . Characteristics of the nets themselves, such as their cost, size, shape, colour, physical condition, type of insecticide used and perceived durability have also been shown to influence net use [7, 19, 20], and are likely to interact with individual, household, community and environmental factors in complex ways to determine attitudes towards net use and the feasibility of net use for a given individual or household.
A recent review of literature on determinants of net use highlights the need for greater understanding of these determinants and the relationships between them . In addition, net use among those who own nets is commonly interrupted by temporary, periodic or infrequent conditions, which can inhibit net use even among regular net users. These conditions include travelling, night work, sleeping in the fields during planting or harvest seasons or while tending livestock, attending late-night social events, disruption of usual sleeping arrangements, net unavailability due to washing or dirtiness, extreme fatigue, labour pains, illness or forgetfulness [13, 27, 30–32].
This study builds on a previous analysis at the net level  of the results of two sequential surveys in Ethiopia  that demonstrated certain modifiable factors concerning net use, such as improved net care and replacement. In the previous study, the outcome was use or non-use for each net , whereas here factors associated with the outcome of individual use or non-use were investigated. The current study adjusts for both household and net level factors, including the important net age and condition characteristics identified previously . In contrast to most previous studies, including some in Ethiopia [19, 33, 34], this analysis was not restricted to high risk groups (children U5 and pregnant women), but examined net use in all age groups including the 5-24 year old school-age children and young adults, women of reproductive age, and adults age 50 years and older. School-age children are a group generally least protected by insecticide-impregnated nets in Africa , including Ethiopia .
Since availability of a net in the household is a prerequisite for use, only those households owning at least one net are included in this analysis, to avoid biasing the results by lack of net availability in some households. Net density (i.e. number of nets per person in each household) is also used to account for differing household sizes and intra-household access to nets. The goal of this study is to identify under-served groups and modifiable factors that could be used to better target efforts to increase net use and, hence, assist in long-term reduction of malaria transmission in Ethiopia.